[RC5] C DES core

C. Jasper Spaans spaans at vvtp.tn.tudelft.nl
Thu Jan 22 23:58:05 EST 1998


On Thu, 22 Jan 1998, Mary Conner wrote:

> Given some of the reactions that I have seen to the released x86 clients,
> releasing a functional but slow C DES core may have backfired. 

I don't know why it should be slow. For example: when this client is coded
in C, and doesn't use black magic, it might be possible to let it be
compiled by the compiler.
This would give a client running at at least 90% of the speed of a 
manually
optimized client. 
Just some quick math: suppose we could exhaust the keyspace with a 
""slow""
client in 22.5 days.
If someone were to write an optimized client, this time would drop to
22.5*0.9 which would be about 20.4 days. 

Which leads me to the following conclusion:

If the client is not ready before the contest starts, so be it, just bring
out a client running at 90% of the max. achievable keyrate.

A delay of 2.1 days in releasing the client will not help anyone to 
complete
the contest faster!

(of course, this 2.1 day figure can be extended to 4.2 days for the $5k 
and
8.4 days for the $2k5 prize - all these times have been exceeded)

My $0.02: *** RELEASE SOME CLIENTS NOW - COMPILE THEM WITH GCC -O3 ***

VrGr,
Jasper

  __ ___ __  C. Jasper Spaans
 /     //__  Luchtvaart- en Ruimtevaarttechniek, T.U. Delft
/__ __/ __/  Aerospace Engineering, Delft University of Technology
             finger spaans at vvtp.tn.tudelft.nl for my public PGP-key

--
To unsubcribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list